## _cr16132 ## Source details **Canonical URL:** [_cr16132](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16132.pdf) ## Other formats - [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16132.pdf.md) - [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16132.pdf.json) --- ### MACRO-FINANCIAL LINKAGES - Financial Conditions Index (FCI) - Construction: - Variables: real interest rate of bank loans; real effective exchange rate (REER); real growth of deposits and credit to the private sector; real housing price index (housing component of CPI). - Built as the sum of the cumulative impulse responses of real GDP to each financial variable using VAR analysis. - Estimated using Ordinary Least Squares on quarterly data from 2001 to 2015. - Key findings: - The estimated FCI points to relatively tight financial conditions in 2015. - Financial conditions tightened between mid-2014 and end-2015, driven mostly by an increase in real interest and exchange rates which more than offset faster deposit growth and higher housing prices. - Drivers: declining inflation and inflation expectations in 2015, declining nominal interest rates, relatively stable nominal FX rate, and favorable global financial conditions to capital inflows. - Result: monetary and financial conditions were dragging GDP growth during 2015 despite recent policy rate cuts. - General Equilibrium Model (FSGM) and simulated scenarios - Model features: - IMF’s Flexible System of Global Models (FSGM); multi-region general equilibrium with Cobb-Douglas production tech; OLG consumption block; risk premia (sovereign, household/firm, firm-only, currency premium); 10-year rates via expectation theory with additional risk premium; short-run exchange rate via uncovered interest parity. - Scenarios: - S1: 100 basis points cut in monetary policy rate (full pass-through and slower 1/3 pass-through). - S2: Currency depreciations of 5 and 20 percent with non-linear market risk premia reactions of 25 and 300 basis points respectively. - S3: Reduction in bank lending margins → market risk premia fall by 100 basis points. - S4: Halving collateral requirements → higher credit growth, partly financed by foreign borrowing. - Linearity note: model symmetric and roughly linear aside from zero lower bound nonlinearities; shocks scalable. - Quantified impacts (scenario highlights) - S1 (100 basis points cut): - Full pass-through: one-year 100 basis points decrease → 0.2 percent GDP increase on impact relative to baseline. - The recent 350 basis points cut could increase GDP in 2016 by almost one percent relative to baseline if fully passed through to lending rates. - Slower pass-through (1/3): the 2016 GDP increase would be halved. - S2 (currency depreciation): - 5 percent depreciation + 25 basis points increase in market risk premia: exports rise enough to slightly increase GDP on impact (export price elasticity assumed 0.58). - 20 percent depreciation + 300 basis points increase in market risk premia: GDP decline of 1.5 percent one year after shock due to drops in household consumption and private investment. - S3 (100 basis points reduction in market risk premia): - Bank spreads about 12 percent; a 100 basis points permanent reduction raises private investment and consumption. - Net effect on GDP: <1 percent increase in year 1 to >2 percent after five years. - S4 (halving collateral requirements): - Effective collateral-to-loan ratios hovered over 250 percent as of 2010. - Permanently halving assumed to increase domestic credit by 8 percent of GDP (half from domestic liquidity, remainder from foreign liabilities). - Output rises: about ½ percent in year 1 to over 1 ½ percent by year 6. - Implications - Financial conditions in 2015 moderately tight, dragging growth despite policy easing. - Monetary easing effective only with sufficient pass-through to lending rates. - Large depreciations can be contractionary if they trigger large increases in market risk premia, amplified by high dollarized credit to un-hedged borrowers. - Structural reforms lowering bank spreads or collateral requirements can produce persistent GDP gains (up to 2 percent medium term from bank efficiency; up to over 1 ½ percent after six years from halving collateral). ### BASELINE MACRO PROJECTIONS & CREDIT GROWTH - Projections and historical references: - Staff 2016 projections consistent with credit growth accelerating to 14 percent in 2016 based on historical distributions conditional on financial outturns in emerging markets. - Staff project credit growth to accelerate to 12 percent in 2016 from 11 percent in 2015 (consistent with return of GDP to potential). - Historical: credit rose from 35 percent of GDP in 2005 to 60 percent in 2015. - Distributional positioning: - Projections for real GDP, investment, private consumption, and employment fall within the 25th and 75th percentile of the respective 2014 conditional distributions. - Inflation note: - Inflation forecast lower than historical patterns would imply for projected credit growth, due to positive terms of trade shock from protracted drop in international oil and commodity prices. - Macro-financial risk assessment: - No evidence of a broad macro-financial risk buildup compared to 2014. - Main risks: market and liquidity conditions, and to a lesser extent monetary and financial conditions, driven by high gross foreign liabilities of banking sector and dollarization of domestic credit. - Government deficit and debt high and increasing; deserve close scrutiny though diluted by otherwise stable financial conditions. - Financial system metrics: - 52 supervised institutions: 15 banks (4 state-owned, 3 private domestics, 8 private foreign), 28 savings and loan cooperatives, 9 other non-bank entities. - Banking sector market share ≈ 80 percent. - Total gross assets 90 percent of GDP; total gross liabilities 86 percent of GDP. - About half of total deposits denominated in foreign currency. - State-owned banks account for about 80 percent of deposits in national currency (linked to explicit unlimited state guarantee on all public bank deposit). - Staff view recent credit trend as “in line with healthy financial deepening.” Staff project credit growth supportive of macro activity in short and medium term (¶19). ### BANK-LEVEL STRESS TESTING: SHOCKS, ASSUMPTIONS, RESULTS - Stress scenarios and calibrations: - Credit risk shocks: - (i) aggregate NPL shock = 8 percent of currently performing loans; - (ii) sectoral shock = 6 and 10 percent of currently performing loans to construction and trade respectively (these sectors >40 percent of total loans). - Assumed 75 percent haircut on collateral; provisioning rates 25 percent (domestic currency NPLs) and 50 percent (foreign currency NPLs). - Interest rate risk: - Flow impact from gap between interest-sensitive assets/liabilities; - Stock impact from bond repricing following nominal interest rate increase of 3.5 percentage points. - FX risk: - Direct FX shock; indirect effect on credit quality following 10 percent nominal depreciation assuming 60 percent of FX loans to un-hedged borrowers become non-performing following 100 percent depreciation benchmark. - In some tests, 14 percent depreciation assumed to yield 6 percent of FX loans becoming non-performing with 50 percent provisioning. - Liquidity stress: - Withdrawals per day: demand deposits 10% domestic / 8% foreign; time deposits 5% domestic / 3% foreign. - Interbank contagion: reverse tests to determine NPL increases required for systemic CAR breaches. - Pre-shock system indicators: - System-wide pre-shock CAR: 17.0 percent (All Banks). - NPL ratio pre-shock: 1.8 percent (All Banks). - ROA (after-tax) pre-shock: 0.7 percent (All Banks). - Main stress-test outcomes: - System-wide post-shock CAR remains above 10 percent even after an extremely negative combined shock. - System could withstand a liquidity drain for 5 days. - Combined shock does not trigger any bank failure, though CAR of nine banks falls below regulatory threshold and would require balance-sheet adjustment. - Indirect FX shock through increased credit risk of un-hedged borrowers is most significant; could reduce CAR of five privately owned and foreign banks below regulatory threshold. - Interbank exposures limited; no domino contagion observed. - Quantified stress impacts (selected figures): - Credit shock (NPLs +8% of performing loans): post-shock CAR All Banks = 15.2 percent; CAR change = -1.7 percentage points. - Sectoral shock: post-shock CAR All Banks = 16.2 percent; CAR change = -0.6 percentage points. - Interest rate shocks: - Net interest income impact: post-shock CAR All Banks = 17.7 percent; CAR change = 0.7 percentage points. - Repricing impact: post-shock CAR All Banks = 16.9 percent; CAR change = -0.8 percentage points. - Overall NII and repricing combined: change = -0.1 percentage points (All Banks). - FX risk: - Direct FX risk: post-shock CAR All Banks = 16.4 percent; CAR change = -0.7 percentage points. - Indirect FX risk: post-shock CAR All Banks = 14.9 percent; CAR change = -1.5 percentage points. - Overall direct and indirect FX change = -2.2 percentage points (All Banks). - Interbank contagion after macroshocks: CAR = 13.5 percent (All Banks); CAR after first iteration = 13.5 percent (no additional contagion). - Liquidity stress test (liquid banks after 5 days): All Banks = 15; State Owned = 4; Domestic Private = 5; Foreign = 3. - Reverse stress test (post-all fundamental shocks): Post-shock CAR All Banks = 13.1 percent; State Owned = 13.9 percent; Domestic Private = 7.8 percent; Foreign = 12.2 percent. - Change in CAR (all fundamental shocks): All Banks = -4.0 percentage points; State Owned = -4.1; Domestic Private = -4.4; Foreign = -3.7. - Post-shock liquidity ratios (after 5 days): - Liquid assets/total assets post-shock: All Banks = 4.0 percent; State Owned = 2.4 percent; Domestic Private = 4.8 percent; Foreign = 7.1 percent. - Liquid assets/short-term liabilities post-shock: All Banks = 36.3 percent; State Owned = 16.7 percent; Domestic Private = 106.9 percent; Foreign = 117.1 percent. - Profit buffer: Profits (10-year average)/pre-shock RWA = 0.4 percent (All Banks); using profits for defense raises post-shock CAR modestly (All Banks CAR if profit used for defense = 13.5 percent). - Policy-relevant conclusions: - Financial sector overall sound and can absorb a range of shocks. - Vulnerabilities: low bank profitability, high FX exposure, widespread dollarization of loans and deposits, high and increasing government deficit and debt. - Policy emphasis: - Monitor government fiscal metrics closely. - Address dollarization vulnerabilities. - Strengthen profitability and buffers to improve resilience. - Maintain vigilance on market and liquidity conditions given banking sector’s gross foreign liabilities and dollarization. ### BANK CONTAGION MODULE: INTERNATIONAL SPILLOVERS TO COSTA RICA - Methodology: - IMF Bank Contagion Module using BIS banking statistics and bank-level data to estimate rollover risks from foreign banks’ affiliates and cross-border lending; assumes uniform deleveraging and no bank recapitalizations or remedial policy actions. - Transmission via Panama incorporated proportionally. - Quantitative spillovers (based on 2015Q3 data): - 10 percent shock to U.S. and Canada assets: reduces credit in Costa Rica by 6.3 percent of GDP (7.9 percent of total domestic and cross-border credit). - 10 percent shock to U.S. only: reduces cross-border credit availability by 2.8 percent of GDP. - 10 percent shock to Canada only: reduces credit by 3.6 percent of GDP. - 10 percent shock to selected large European countries: reduces credit by about 2.4 percent of GDP (0.6 percent of GDP transmitted via Panama). - Regional comparisons: similar 10 percent shock in U.S. and Canada reduces credit in Guatemala and Honduras by 2.7 and 1.4 percent of GDP respectively. - Upper-limit rollover risk: about 11.6 percent of GDP (14.6 percent of total domestic and cross-border credit) in a worst-case no-replacement scenario. - Notes: - Simulations based on 2015Q3 data. - Advanced economy banking systems required to hold 8.5% capital ratio in these estimates (higher than prior 6%), increasing deleveraging for given shocks. ### MONETARY POLICY STANCE ASSESSMENT - Conclusion summary: - Monetary policy appropriately expansionary after eight consecutive policy rate decreases since early 2015. - Current nominal policy rate: 1.75 percent (as of January 4, 2016). - This is below the estimated neutral monetary policy rate; BCCR should be ready to reverse easing if price pressures materialize. - Drivers of 2015 disinflation: fall in international commodity prices and a widening output gap. - Inflation target range decreased from 3–5 percent to 2–4 percent. - Neutral interest rate estimation (five approaches) - UIP condition (Mar 2006–Nov 2015): Neutral nominal interest rate = 4.6 percent. - Expected-inflation augmented Taylor rule (Mar 2006–Nov 2015): Neutral real = 2.4 (real neutral rate); Neutral nominal = 5.4 percent (staff projected inflation of 3 percent end-2016). - General equilibrium model: Nominal neutral rate = 5.6 percent. - Forward-looking monetary model (Jan 1994–Aug 2015): Nominal neutral = 4.8 percent. - QPM (1996Q1–2016Q1): Nominal neutral = 4.4 percent. - Average nominal neutral across five models = 4.6 percent (2014 estimate was 5.1 percent). - Policy guidance: - Actual nominal policy rate = 1.75 percent; stance appropriately expansionary given positive output gap and inflation projected below target in near term. - Risks warranting tightening: - (i) faster U.S. growth; - (ii) rebound in international commodity prices; - (iii) upward food price pressures from regional droughts; - (iv) second-round effects from currency depreciation following global interest rate normalization. ### FINANCIAL DEEPENING: FINDINGS, POTENTIAL, AND RECOMMENDATIONS - Current state: - Financial system deepened notably over past decade driven by institutional growth; market development stagnated. - Costa Rica lags other emerging markets on most financial market development subcomponents; institutional access high (ATMs, branches per 100,000 adults) but institutional efficiency low (high spreads, overhead costs, net interest margins). - Institutional depth low: low credit and deposits to GDP; small mutual fund and insurance sectors. - Financial development below levels predicted by fundamentals on most subcomponents; weak legal collateral framework increased collateral-to-loan ratio historically. - Relationship with growth and stability: - Non-linear (inverted U-shape) relationship between financial development and growth; access and efficiency produce unambiguously positive growth effects but may raise stability costs if profitability drops and risk-taking rises. - Costa Rica has not reached levels that maximize growth–stability benefits. - Policy recommendations: - Short-term: remove distortions preventing full financial development given current fundamentals. - Modernize secured transactions and collateral framework (note: 2015 secured transactions law adopted; broadened assets usable as collateral and allowed out-of-court enforcement; careful monitoring warranted). - Improve intermediation efficiency and level playing field between private and public banks; remove explicit state guarantee on colon-denominated deposits in state banks. - Market-friendly debt management to foster secondary markets: standardized instruments, conventional maturities, strengthen legal/regulatory frameworks. - Promote stock market development: macro stability, stronger institutional and legal frameworks for investor rights and disclosure, policies to increase market size and institutional investor base; encourage private entry in insurance sector; strengthen minority investor protection; review tax treatment of securities issuance/investment with fiscal sustainability in mind. - Long-term: further financial development beneficial if accompanied by adequate regulatory oversight to prevent excesses. ### FINANCIAL INCLUSION: HOUSEHOLD VS FIRM INCLUSION - Overview: - Household inclusion strong—Costa Rica a champion among LAC on household inclusion indices (account holdings, ATM usage, debit cards). - Enterprise inclusion weak—high collateral requirements, cost/access to finance are major SME constraints; weak access to physical infrastructure; mobile banking and correspondent banking underperform peers. - Costa Rica scores below emerging market average on Global Microscope; strengths in prudential regulation and credit reporting, weaknesses across other Microscope dimensions relevant for enterprise inclusion. - Financial inclusion gaps: - Households: positive gap overall (ATMs and debit card usage offset negative gaps on savings/borrowing and credit card usage). - Firms: negative gap driven by high collateral, access/cost constraints, low SME usage of banks. - Determinants associated with higher inclusion: higher income per capita, higher education (households), stronger rule of law, lower informality, lower share of foreign-owned firms. - Econometric evidence and implications: - Regression associations: higher (less negative) inclusion gaps linked to lower non-interest income (household & firms), lower bank safety buffers (households), lower bank overhead costs (firms), stronger Global Microscope score (firms). - Policy implication: strengthen regulatory environment for financial inclusion (Microscope dimensions); formalize micro-lenders/savers; adopt explicit regulations for banking agents, e-money, micro-insurance. - Micro-founded model experiments (three simulations) - (i) reduce participation costs to 0; - (ii) relax collateral to world minimum; - (iii) reduce intermediation monitoring costs to 0 (spreads = NPL). - Results: - Removing participation costs benefits the excluded and reduces inequality. - Relaxing collateral yields largest GDP gains but may increase inequality if benefits accrue mainly to already-credit-accessible high-productivity firms; can reduce or raise inequality depending on distributional uptake. - Reducing intermediation costs raises GDP but may moderately increase inequality as benefits concentrate on already-credited firms. - Key statistics: - Share of firms with access to credit near 60 percent. - Collateral needed in Costa Rica compared to peers and spreads: interest rate spreads in Costa Rica are 2.5 times higher than in Panama and El Salvador. - Policy recommendations: - Modernize collateral framework and strengthen regulatory environment for financial inclusion. - Ensure level playing field between private and public banks to lower intermediation costs, acknowledging potential distributional trade-offs. - Complementary measures: formalize micro-lenders, strengthen regulation of banking agents and e-money. ### RELAXING COLLATERAL CONSTRAINTS — GROWTH AND INEQUALITY TRADEOFFS - Model findings: - High collateral requirements constrain many firms; relaxing collateral raises growth and can lower inequality by enabling medium-sized firms to expand and raise wages. - Reducing monitoring/intermediation costs yields high growth but may increase inequality modestly (benefits concentrate on firms already with credit). - Mechanisms: collateral reforms affect both extensive (new firms access) and intensive margins (higher borrowing for existing firms); intermediation cost reductions primarily affect intensive margin. - Calibration and data: - Baseline calibrated to 12 LAC countries; firm-level data from 2005 Enterprise Survey; focuses on formal sector firms. - Figures (authors’ calculations) indicate Costa Rica benefits in GDP and GINI from reducing collateral and participation costs; reducing monitoring costs increases GDP and may raise GINI modestly. - Policy emphasis: - Modernize collateral framework to raise growth and inclusion while monitoring distributional impacts. - Strengthen regulatory environment and level playing field to lower intermediation costs cautiously. ### FISCAL SUSTAINABILITY: DEFICITS, DEBT, AND ADJUSTMENT NEEDS - Recent fiscal developments: - Primary and overall deficits peaked at 3 and 5½ percent of GDP by 2010. - Central government primary deficit broadly flat at 3 percent of GDP in 2014; public debt rose to 42½ percent of GDP in 2015 (from 24 percent in 2008). - Interest bill increased moderately; overall deficit ≈ 6 percent of GDP in recent years. - CCSS surplus 0.75 percent of GDP in 2015; public sector enterprises small deficit 0.1 percent of GDP; central bank operational deficit about 0.75 percent of GDP. - Required adjustment and scenarios: - Total adjustment of about 3¾ percent of GDP necessary to restore debt sustainability. - Passive scenario: primary deficit projected to rise to 3¼ percent of GDP in 2016 and to 3½ over medium term as constitutionally-mandated education spending reaches 8 percent of GDP. - Debt-stabilizing primary surplus needed = ¼ percent of GDP given projected increase in real rates above real GDP growth. - Timing matters: postponing retrenchment raises required improvement. - Optimal consolidation: gradual but frontloaded; staff recommend spreading adjustment into second and third years with full sustainability gap closed by end of 5-year projection. - Key fiscal parameters (Fund staff estimates) - Medium-term primary balance (I): -3.5 - 2015 outturn: -3.0 - 2016 projected change: -0.2 - 2017–21 projected change: -0.2 - Debt stabilizing primary balance (II): 0.3 - 2021 real growth: 4.0 (percent) - 2021 real interest rate: 4.6 (percent) - 2021 debt: 55.5 - Fiscal Sustainability Gap (II − I): 3.8 - Pension actuarial considerations: - Including public pension actuarial deficit increases required adjustment by additional 1½ percent of GDP to achieve actuarial balance over next 50 years; actuarial gap almost 4 percent of GDP over 100 years. - CCSS covers ⅔ of labor force (~1½ million) with ~165,000 beneficiaries; legacy regimes financed by central government transfers ≈ 2½ percent of GDP in 2015. - Authorities’ measures and staff assessment: - Authorities submitted measures expected to yield ~2¾ percent of GDP (revenue measures ~slightly more than 2 percent of GDP; expenditure provisions reduce outlays by ¼ percent of GDP in 2016). - VAT reform: broaden base to services; gradual rate increase from 13 to 15 percent starting in 2016; targeted reductions in basic goods basket conditional on transfers. - Income tax reform: two higher marginal rates 20 and 25 percent; unify capital income taxation at 15 percent; tax capital gains. - Staff baseline assumes only higher-probability measures; if VAT increases not approved, baseline smaller. - Consolidation matrix (selected entries, percent of GDP) - Total adjustment (Authorities' plans): 4.0 - Baseline Scenario (I): 2.9 - Additional Measures (II): 2.3 - Total adjustment (I + II): 1.5 - Staff Fiscal Sustainability Gap: 3.8 - Revenue items (Authorities | Baseline | Additional | etc.): - Revenue total: 2.7 | 2.5 | 1.7 | 0.8 | 2.5 - VAT: 1.3 | 1.3 | 0.5 | 0.8 | 1.3 - Income tax: 0.6 | 0.6 | 0.6 | -0.6 - Expenditure total: 1.3 | 0.4 | 0.6 | 0.7 | 1.3 - Notes: staff only include higher-probability revenue measures; baseline assumes VAT rate increases will not be approved by Congress. - Policy guidance: - Consolidation on both revenue and expenditure sides; revenue mobilization cornerstone. - Emphasize progressive revenue measures protecting low-income households via targeted transfers. - Political agreement crucial; full implementation permits balanced macro mix and reduces market-rate pressures associated with US normalization. ### DEBT DYNAMICS, RISKS, AND SENSITIVITIES - Shock sensitivities: - A plausible DSA macro-fiscal shock could raise central government debt above 65 percent of GDP by 2021 (>10 percentage points above baseline). - Isolated one standard deviation growth shock in 2016–17 → debt +5 percent of GDP by 2021 relative to baseline. - Fiscal shock: additional 1¼ percent of GDP primary deficit in 2016–17 → debt +3 percent of GDP by 2021. - Interest rate shock: 200 basis points increase → debt +<2 percentage points by 2021. - Exchange rate shock: 15 percent nominal depreciation → debt impact <1 percent of GDP by 2021. - Sovereign financing risks: - Average projected financing needs ≈ 10 percent in 2015–16 (among highest of CAPDR and LA-5). - Spreads on external sovereign bonds increased substantially; Costa Rica lost investment grade rating in 2014 and remains one to two notches below investment grade. - Upside risks to funding costs from Congressional uncertainty on external issuance and global rate normalization; risk of debt spiral if debt rises unabated. - Domestic investor base and bank exposures: - Domestic institutional holdings of domestic government bonds declined from ~80 percent in 2008 to ~60 percent in 2014. - Banks could be close to failing regulatory capital requirements from mark-to-market losses on sovereign exposures under passive no-consolidation scenarios, especially with higher government financing costs. ### BALANCE SHEET & REAL SECTOR HIGHLIGHTS - Net external positions: - Total net external debtor position ≈ -42 percent of GDP in 2014 (from -35 percent in 2013). - Excluding FDI, small net creditor position ≈ 3 percent of GDP in 2014 (down from 4 percent in 2013 and 8 percent in 2010). - Credit and leverage trends: - Bank credit to private sector: <30 percent of GDP in 2004 → ≈55 percent in 2015. - Credit to non-financial corporates: 11 → 17 percent of GDP (2004→2015). - Credit to households: 20 → 35 percent of GDP (2004→2015); bulk increase in domestic currency. - FX credit shares: household FX share ≈30 percent since 2009; corporate FX share ≈65 percent of corporate credit. - Potential output and output gap: - Staff estimate potential output growth in 2015 ≈ 4.0 percent with output gap somewhat negative. - Range of 2015 potential growth across methods: 3.6 to 4.6 percent; average of all models = 3.97 percent. - Output gap estimates vary; Macroframework output gap 2015 = -1.11 percent. - Intel exit effect: - Intel withdrawal lowered actual and potential real GDP by about ¾ percentage points over 2014–2015. - Level effect on potential output: 0.2 percent of GDP in 2014 and 0.8 percent in 2015. ### TOTAL FACTOR PRODUCTIVITY (TFP) — MEASUREMENT CAVEATS - TFP measurement caveats: - TFP is residual; measurement errors in labor/capital attributable to TFP. - Distortions include skilled labor migration, mismeasured input quality, capital utilization changes, and land use. - Indicators consistent with lower measured productivity: - Informality: informal employment increased 40 percent since 2011 while formal sector employment unchanged. - R&D spending < ½ percent of GDP (OECD average ≈ 2.5 percent). - Education and infrastructure outcomes weak relative to spending: education spending ≈ 7–8 percent of GDP but outcomes similar to Chile which spends ~50 percent less; infrastructure quality among worst in OECD per WEF. ### CROSS-COUNTRY REGRESSIONS: FEMALE LABOR FORCE PARTICIPATION (FLFP) AND DEVELOPMENT DRIVERS - Key regression findings: - Investment in transport and telecommunications and internet access positively associated with female LFP. - Female tertiary enrollment positively and significantly associated with female LFP; male tertiary attainment shows negative coefficients in some specifications. - Fertility rates have negative and marginally significant coefficients; Costa Rica fertility = 1.8 children per woman. - U-shaped relationship between female LFP and log GDP per capita; Costa Rica located near the trough. - Decomposition relative to LA5: - Differences in investment (transport & telecoms) explain a large portion of Costa Rica’s lower female LFP vs LA5. - Policy recommendations: - Increase investment in physical infrastructure and ICT. - Improve implementation capacity to remove investment constraints. - Explore measures to support working mothers and understand childcare/fertility interactions. - Selected regression statistics (preserved) - Log GDP per capita coefficients across models: e.g., -46.918*** (11.909) through -113.19*** (32.321). - (Log GDP per capita)^2 coefficients: e.g., 2.511*** (0.641) through 5.963*** (1.857). - Observations and R-squared vary by model (examples: observations 4073 with R squared 0.495; 303 with R squared 0.710). *Italic: Source — IMF staff report excerpts and simulations contained in content unit _cr16132.* ### References ____________________________________________________________________________________ 10 ### MACRO-FINANCIAL LINKAGES ### Financial Conditions Index (FCI) - Purpose and construction: - Summarizes information in key financial variables and captures correlation with economic activity. - Financial variables included: - the real interest rate of bank loans; - the real effective exchange rate (REER); - the real growth of deposits and credit to the private sector; - a real housing price index (approximated by the housing component of the consumer price index). - Built as the sum of the cumulative impulse responses of real GDP to each financial variable using VAR analysis. - Estimated using Ordinary Least Squares on quarterly data from 2001 to 2015. - Key empirical finding: - The estimated FCI points to relatively tight financial conditions in 2015. - Financial conditions turned tighter between mid-2014 and end-2015, driven mostly by an increase in real interest and exchange rates which more than offset faster deposit growth and higher housing prices. - The increase in real interest and exchange rates is related to declining inflation and inflation expectations in 2015, declining nominal interest rates, and a relatively stable nominal FX rate, while global financial conditions remained favorable to capital inflows. - Result: monetary and financial conditions were dragging GDP growth during 2015 despite recent policy rate cuts; contrast with the propelling cycle in 2012–13 and neutral financial conditions in the first half of 2014. ### General Equilibrium Model Simulations (FSGM) - Model description: - Used the IMF’s Flexible System of Global Models (FSGM) developed by the Economic Modeling Division of IMF’s Research department. - Multi-region, general equilibrium system combining micro-founded and reduced-form relationships. - Supply side pinned down by Cobb-Douglas production technology. - International linkages modeled in aggregate by country/region; public debt levels influence global real interest rate via national savings. - Parameters largely estimated from data; OLG consumption block gives non-Ricardian properties with endogenous national savings; government absorption exogenous. - Transmission of financial shocks occurs through changes in interest rates and risk premia. - Model includes risk premia: sovereign (applies to entire domestic economy), a premium applying to households and firms, one applying only to firms, and a currency (country) premium. - 10-year interest rates determined by expectation theory of term structure with additional risk premium; interest rates for consumption, investment, government debt, and net foreign assets are weighted averages of 1- and 10-year nominal rates. - Short-run exchange rate determined via uncovered interest parity; long-run adjusts to ensure external stability given households’ desired net foreign asset holdings. - Scenarios simulated (four): - S1: Expansionary monetary policy shock — a 100 basis points cut in the monetary policy rate, modeled under full pass-through and under a slower, more persistent transmission to lending rates. - S2: Currency depreciations of 5 and 20 percent that include a non-linear reaction in market risk premia of 25 and 300 basis points respectively. - S3: Reduction in bank lending margins resulting in a reduction in market risk premia of 100 basis points. - S4: Decrease in collateral requirements yielding higher credit growth partially financed by higher foreign borrowing. - Note: aside from zero lower bound nonlinearities, the model is symmetric and roughly linear; shocks can be scaled. ### Scenario Results and Quantified Impacts - S1: Monetary policy cut (100 basis points) - If fully passed through to domestic lending rates: - A one-year 100 basis points decrease in the monetary policy rate results in a 0.2 percent GDP increase on impact in the model compared to the baseline. - The recent 350 basis points cut in the monetary policy rate could increase GDP in 2016 by almost one percent relative to the baseline, if fully passed through to domestic lending rates. - Channels: higher domestic absorption and export demand induced by colon depreciation following lower real interest rates compared to foreign rates. - Under slower pass-through (only 1/3 of impulse passed to lending rates): - The resulting increase in GDP in 2016 would be halved. - S2: Currency depreciation with nonlinear market premia reaction - Under a 5 percent depreciation with a 25 basis points increase in market risk premia: - Exports increase sufficiently to offset adverse impact on domestic demand from higher risk premia, resulting in a slight increase in GDP on impact. - Assumption: a 0.58 price elasticity of exports is used in the simulations. - Under a 20 percent depreciation with a 300 basis points increase in market risk premia: - The increase in risk premia prevails, inducing a GDP decline of 1.5 percent one year after the shock, as decreases in household consumption and private investment more than offset the boost from higher exports. - Summary: a large currency depreciation could decrease output by up to 1.5 percent in the short and medium term via higher NPLs and decreased risk appetite, especially given high dollarized credit to un-hedged borrowers. - S3: Reduction in market risk premium (100 basis points) - At about 12 percent, bank spreads are exceptionally high in Costa Rica and constrain access to credit. - A 100 basis points permanent reduction in market risk premia (reflecting decreased bank lending margins) increases private investment and household consumption. - Resulting effects: - Domestic absorption rises; inflation increases; policy rate increases; colon appreciates and exports decline. - Net effect on GDP is positive and persistent: - Increase ranges from less than 1 percent in the first year to more than 2 percent after five years. - Conclusion: increasing efficiency and competition in the banking sector could spur GDP by up to 2 percent in the medium term. - S4: Halving collateral requirements - Effective collateral-to-loan ratios were reported hovering over 250 percent as of 2010. - Permanently halving collateral requirements assumed to increase domestic credit by 8 percent of GDP, about half financed by domestic excess liquidity and the remainder by increased foreign liabilities. - Effects: - Increases domestic absorption, slightly depresses exports, raises private investment and the stock of private capital, contributing to persistent increases in actual and potential output. - Output rises above potential opening a moderate positive output gap. - Overall medium term GDP increase ranges from ½ percent in the first year to over 1 ½ percent in the sixth year. ### Summary Findings and Implications - Financial conditions in 2015 were moderately tight and were a drag on GDP despite policy easing. - Monetary easing can materially boost GDP if passed through to lending rates; incomplete pass-through substantially reduces the effect. - Large colon depreciations can be contractionary when they provoke large increases in market risk premia due to loan dollarization and higher NPLs. - Structural reforms that lower bank spreads or collateral requirements could produce persistent and sizable GDP gains: - Bank efficiency/competition could raise GDP by up to 2 percent in the medium term. - Halving collateral requirements could raise GDP by up to over 1 ½ percent after six years. - Baseline projections show no evidence of macro-financial risk buildup and credit growth remains supportive of macroeconomic activity (as summarized in the opening section). *Source: IMF staff estimates and simulations as presented in "MACRO-FINANCIAL LINKAGES" (Prepared by Valentina Flamini), IMF country report content.* ### 10. Baseline macro projections are consistent with accelerating credit growth in the short ### 10. Baseline macro projections are consistent with accelerating credit growth in the short term ### Baseline macro projections and credit growth - Based on the historical distribution of macro outcomes conditional on financial outturns in emerging markets, the team’s 2016 macroeconomic projections are consistent with credit growth accelerating to 14 percent in 2016. - Staff projections (consistent with return of GDP to its potential) indicate credit growth to accelerate to 12 percent in 2016 from 11 percent in 2015. - Historical trend: credit rose from 35 percent of GDP in 2005 to 60 percent in 2015. - The projections for real GDP, investment, private consumption, and employment growth fall within the 25th and 75th percentile of the respective 2014 conditional distributions. - Inflation forecast is lower than historical patterns would imply for the projected credit growth, driven by a positive terms of trade shock induced by the protracted drop in international oil and commodity prices. ### Assessment of macro-financial risks and vulnerabilities - No evidence of a broad macro-financial risks buildup compared to 2014. - Market and liquidity conditions, and to a lesser extent monetary and financial conditions, pose the most significant risks due to: - high gross foreign liabilities of the banking sector, and - dollarization of domestic credit. - Government deficit and debt are high and increasing and “deserve close scrutiny,” but related risks are diluted by otherwise stable financial conditions. - Persistent vulnerabilities highlighted: - Heavy dollarization of credit and deposits. - Low bank profitability and high FX exposure intensify vulnerability to external shocks. ### Financial system structure and trends (key facts and metrics) - Financial system composition: - 52 supervised institutions, of which 15 banks (4 state-owned, 3 private domestics, and 8 private foreign), 28 savings and loan cooperatives, and 9 other non-bank financial entities. - Banking sector retains about 80 percent of the market share. - Total gross assets and liabilities of the system amount to 90 and 86 percent of GDP respectively. - Deposit and currency structure: - About half of total deposits is denominated in foreign currency. - State-owned banks account for about 80 percent of deposits in national currency (linked to explicit unlimited state guarantee on all public bank deposit). - Financial deepening: credit growth in both national and foreign currency has exceeded GDP growth over the past decade; staff view recent trend as “in line with healthy financial deepening.” - Specific staff projection and historical reference: - Staff project credit growth to remain supportive of macroeconomic activity in the short and medium term (¶19). ### Bank-level stress testing: shocks, assumptions, and calibrations - Credit risk shocks: - (i) aggregate NPL shock equivalent to 8 percent of currently performing loans; - (ii) sectoral shock equivalent to 6 and 10 percent of currently performing loans to the construction and trade sectors respectively (together account for more than 40 percent of total loans). - Assumed a 75 percent haircut on collateral, and a 25 and 50 percent provisioning rate on new NPLs in domestic and foreign currency respectively (footnote). - Interest rate risk shocks: - (i) flow impact from gap between interest sensitive assets and liabilities; - (ii) stock impact from bond repricing following a nominal interest rate increase of 3.5 percentage points. - FX risk shocks: - (i) direct exchange rate risk shock; - (ii) indirect effect on credit quality following a 10 percent nominal depreciation, assuming that 60 percent (the share of FX loans to un-hedged borrowers) of currently performing FX loans would become non-performing following a 100 percent depreciation (benchmark used to calibrate sensitivity). - In some test descriptions, a 14 percent depreciation is assumed to yield 6 percent of FX loans becoming non-performing with a 50 percent provisioning rate. - Liquidity stress tests: - 10 and 8 percent per day withdrawal of demand deposits in domestic and foreign currency respectively; - 5 and 3 percent per day withdrawal of time deposit in domestic and foreign currency respectively. - Interbank contagion: uses bank exposures to illustrate second-round effects; reverse tests determine NPL increases required for (i) system-wide CAR < 10 percent; (ii) at least 8 banks to fall below regulatory CAR; (iii) 50 percent of total bank market share to fall below regulatory CAR. ### Stress test outcomes and resilience - Pre-shock system indicators: - System-wide pre-shock CAR: 17.0 percent (All Banks). - NPL ratio pre-shock: 1.8 percent (All Banks). - ROA (after-tax) pre-shock: 0.7 percent (All Banks). - Main stress-test findings: - The system-wide post-shock CAR remains above 10 percent even after an extremely negative combined shock (aggregate credit shock, increase in interest rates, and FX depreciation). - The system could withstand a liquidity drain for 5 days. - Combined shock does not trigger any bank failure, although CAR of nine banks falls below the regulatory threshold and would require balance-sheet adjustment. - Indirect FX shock through increased credit risk of un-hedged borrowers is the most significant: could reduce CAR of five privately owned and foreign banks below the regulatory threshold given higher exposure to dollarized loans. - Interbank exposures are limited; no domino contagion effects observed. - Quantified stress impacts (Table 4 / Table 5 highlights): - Credit shock (proportional increase in NPLs of 8 percent of performing loans): post-shock CAR All Banks = 15.2 percent; CAR change = -1.7 percentage points. - Sectoral shock: post-shock CAR All Banks = 16.2 percent; CAR change = -0.6 percentage points. - Interest rate shocks: - Net interest income impact: post-shock CAR All Banks = 17.7 percent; CAR change = 0.7 percentage points. - Repricing impact: post-shock CAR All Banks = 16.9 percent; CAR change = -0.8 percentage points. - Overall NII and repricing combined: change = -0.1 percentage points (All Banks). - FX risk: - Direct FX risk: post-shock CAR All Banks = 16.4 percent; CAR change = -0.7 percentage points. - Indirect FX risk: post-shock CAR All Banks = 14.9 percent; CAR change = -1.5 percentage points. - Overall direct and indirect FX change = -2.2 percentage points (All Banks). - Interbank contagion after macroshocks: CAR = 13.5 percent (All Banks); CAR after first iteration = 13.5 percent (no additional contagion). - Liquidity stress test (number of liquid banks after 5 days): All Banks = 15; State Owned = 4; Domestic Private = 5; Foreign = 3. - Reverse stress test (Table 5 summary): - Post-shock CAR (percent of post-shock RWA) after all fundamental shocks: All Banks = 13.1 percent; State Owned = 13.9 percent; Domestic Private = 7.8 percent; Foreign = 12.2 percent. - Change in CAR (all fundamental shocks) All Banks = -4.0 percentage points; State Owned = -4.1; Domestic Private = -4.4; Foreign = -3.7. - Post-shock liquidity ratios (after 5 days): - Liquid assets/total assets post-shock: All Banks = 4.0 percent; State Owned = 2.4 percent; Domestic Private = 4.8 percent; Foreign = 7.1 percent. - Liquid assets/short-term liabilities post-shock: All Banks = 36.3 percent; State Owned = 16.7 percent; Domestic Private = 106.9 percent; Foreign = 117.1 percent. - Profit buffer: - Profits (10-year average)/pre-shock RWA = 0.4 percent (All Banks); using profits for defense raises post-shock CAR modestly (e.g., All Banks CAR if profit used for defense = 13.5 percent). ### Key policy-relevant conclusions and recommendations - Financial sector is overall sound and could absorb a range of shocks, but: - Low bank profitability and high FX exposure increase vulnerability to external shocks. - Widespread dollarization of loans and deposits remains a persistent vulnerability and amplifies credit growth volatility through exchange rate pass-through. - High and increasing government deficit and debt require close scrutiny despite currently diluted overall macro-financial risks. - Policy emphasis implied by analysis: - Monitor government fiscal metrics closely to prevent amplification of systemic risks. - Address dollarization vulnerabilities (given high FX loans/total loans and net FX exposures). - Strengthen profitability and buffers to improve resilience to prolonged adverse shocks. - Maintain vigilance on market and liquidity conditions given banking sector’s gross foreign liabilities and dollarization. *Italic: Source — IMF staff report excerpt: Costa Rica financial stability analysis and stress-testing results (content unit _cr16132).* ### 8. We used the IMF Bank Contagion Module to assess the impact of financial spillovers to ### 8. We used the IMF Bank Contagion Module to assess the impact of financial spillovers to Costa Rica from stress in international banks ### Bank contagion analysis and methodology - Model: IMF Bank Contagion Module using BIS banking statistics and bank-level data; estimates potential rollover risks for Costa Rica from (i) foreign banks’ affiliates operating in Costa Rica and (ii) foreign banks’ direct cross-border lending to Costa Rica borrowers. - Rollover-triggering assumption: bank losses in the value of private and public sector assets in certain countries and/or regions. If banks lack sufficient capital buffers to cover losses in a scenario, they deleverage to restore capital-to-asset ratios, reducing foreign and domestic assets and squeezing credit lines to Costa Rica and other countries. - Transmission via Panama: estimated impact on cross-border credit availability for Costa Rica incorporates transmission of shocks through Panama. Assumption: cross-border lending from Panama to Costa Rica declines proportionally to the decline in cross-border lending to Panama from the banking systems where the shocks originate. - Bank recapitalizations and remedial policy actions (e.g., ring fencing, monetary policy) at host and/or home country level are not assumed. ### Key quantitative findings on spillovers (as of October 2015) - Impact of a 10 percent shock to assets originating in the U.S. and Canada: - Reduce credit in Costa Rica by 6.3 percent of GDP. - Equivalent to 7.9 percent of total domestic and cross-border credit to the public and private sectors. - Impact of a 10 percent shock to U.S. assets only: - Reduce cross-border credit availability in Costa Rica by 2.8 percent of GDP. - Impact of a 10 percent shock to Canadian assets only: - Reduce credit in Costa Rica by 3.6 percent of GDP. - Impact of a 10 percent shock to European assets (selected large European countries aggregated): - Reduce credit availability to Costa Rica by about 2.4 percent of GDP. - Almost one third of this estimated credit loss (0.6 percent of GDP) from a Europe-origin shock would be transmitted through cross-border lending from Panama. - Regional comparisons: - A similar 10 percent shock in the U.S. and Canada would reduce credit in Guatemala and Honduras by 2.7 and 1.4 percent of GDP respectively. - Spillovers to Costa Rica from stress in international banks are larger than in regional peers, except for Panama and El Salvador. - Upper-limit (worst-case) rollover risk: - Upper limit of rollover risks on external credit for Costa Rica is about 11.6 percent of GDP. - Equivalent to 14.6 percent of total domestic and cross-border credit to the public and private sectors in Costa Rica. - This upper limit corresponds to a worst case scenario without any replacement, either domestic or external, of the loss of credit by BIS reporting banks to Costa Rican borrowers. ### Additional methodological/contextual notes - Simulations are based on 2015Q3 data. - Percent reductions reported are reductions in foreign banks' credit due to the impact of the shock on their balance sheet, assuming uniform deleveraging across domestic and external claims. - Consolidated claims on Costa Rica of BIS reporting banks are used—excluding domestic deposits of subsidiaries of these banks in Costa Rica. - Total credit to the non-bank sectors in Costa Rica is calculated by adding IFS local banks’ claims on non-bank borrowers and BIS reporting banks’ direct cross-border claims on non-bank sectors (BIS Locational Banking Statistics Table 6B). - Change in simulation outcomes relative to earlier estimates partly reflects the requirement that advanced economy banking systems hold an 8.5% capital ratio (in line with Basel III) to be considered “adequately capitalized,” compared with 6% in previous estimates; higher required minimum capital leads to greater deleveraging for any given shock. *Italic: Source: Research Department Macro-Financial Division Bank Contagion Module based on BIS, ECB, IFS, and Bankscope data (simulations based on 2015Q3 data).* --- ### Monetary policy stance assessment ### Purpose and approach - Objective: empirical assessment of the current monetary policy stance in Costa Rica. - Conclusion summary: monetary policy is appropriately expansionary after eight consecutive policy rate decreases since early 2015; the current nominal monetary policy interest rate of 1.75 percent is below the estimated neutral monetary policy rate. The BCCR should be ready to reverse easing if price pressures materialize. - Drivers of 2015 disinflation: fall in international commodity prices (lower domestic prices of regulated goods and services, and agricultural products) and a widening output gap. - Central bank actions: - Cut the monetary policy rate by 350 cumulative basis points since the beginning of 2015, to 1.75 percent in January 4, 2016. - Decreased the inflation target range from 3–5 percent to 2–4 percent. ### Estimation strategy for neutral interest rate (NIR) - Five empirical approaches used: 1. Uncovered Interest Parity (UIP) condition (monthly data March 2006 to November 2015). 2. Expected-inflation augmented Taylor rule (monthly data March 2006 to November 2015). 3. General equilibrium model focusing on aggregate demand and supply (monthly data March 2006 to November 2015). 4. Forward-looking monetary model à la Clarida, Galí, and Gertler (data January 1994 to August 2015). 5. Linear semi-structural new-Keynesian Quarterly Projection Model (QPM) with model-consistent expectations (quarterly data 1996Q1 to 2016Q1). - Models estimate short-run operational neutral interest rate and compare it with the actual policy rate to assess stance. ### Model-specific neutral rate estimates and relevant numbers - UIP condition: - Neutral nominal interest rate: 4.6 percent. - Model assumes implicit annual nominal depreciation in line with inflation differential with the U.S. to maintain constant real exchange rate, plus a country risk premium. - Expected-inflation augmented Taylor rule: - Neutral real level: 2.4 (real neutral rate). - Neutral nominal interest rate: 5.4 percent (with staff’s projected inflation of 3 percent at the end 2016). - Note: estimated transmission from monetary policy rates to inflation is 5 to 6 months in Costa Rica. - General equilibrium model: - Nominal neutral rate: 5.6 percent. - Forward-looking monetary model: - Nominal neutral rate: 4.8 percent. - QPM model: - Nominal neutral rate: 4.4 percent. ### Aggregated assessment and policy guidance - Average nominal neutral interest rate across the five models: 4.6 percent. - Comparison with earlier estimate: - 2014 estimate: 5.1 percent. - Decline to 4.6 percent consistent with comparatively lower estimates of potential output and inflation. - Current policy setting: - Actual nominal monetary policy interest rate: 1.75 percent. - Policy stance: appropriately expansionary given a positive output gap and inflation projected to remain below target in the near term. - Risks that would warrant tightening (BCCR should stand ready to raise interest rates if these materialize): - (i) faster U.S. growth; - (ii) a rebound in international commodity prices; - (iii) upward food price pressures stemming from regional droughts; - (iv) possible second-round effects from currency depreciation following the normalization of global interest rates. *Italic: Source: IMF staff assessment and model estimates (various models and data periods as detailed in the text).* ### References ### References ### Bibliographic citations - Berg, A., P. Karam, and D. Laxton, 2006, “A Practical Model-Based Approach to Monetary Policy Analysis—Overview,” Working Paper 06/80 (Washington: International Monetary Fund). - Blanchard, O. J., and C. M. Kahn, 1980, “The Solution of Linear Difference Models under Rational Expectations,” Econometrica, 48, pp. 1305-1311. - Clarida, R., J. Galí and M. Gertler, 1998, “Monetary Policy Rules in Practice Some International Evidence,” European Economic Review, 42, pp. 1033-1067. - Durbin, J., and S. J. Koopman (2001). Time Series Analysis by State Space Methods. Oxford University Press, Oxford. - Galí, J and T. Monacelli, 2005, “Monetary Policy and Exchange Rate Volatility in a Small Open Economy,” Review of Economic Studies, 72, pp. 707-734. - Hamilton, J., (1995). Time Series Analysis. Princeton, NJ: Princeton University Press. - Magud, N., and E. Tsounta, 2012, “To Cut or Not to Cut? That is the (Central Bank’s) Question: In Search of the Neutral Interest Rate in Latin America,” Working Paper 12/243 (Washington: International Monetary Fund). ### Financial Deepening in Costa Rica — Major findings - The note examines the current state of financial development in Costa Rica and implications for potential growth and stability from further financial deepening. - Costa Rica’s financial system continues to lag behind those of other emerging markets as well as the country’s macroeconomic fundamentals. - Short-run focus: remove distortions that prevent Costa Rica from reaching its full financial development potential given current macroeconomic fundamentals. - Long-run: further financial development would be beneficial for growth and stability, provided adequate regulatory oversight prevents excesses. ### A. Financial Development: Where Costa Rica stands - Assessment method: - A comprehensive index covering 123 countries for the period 1995–2013 (see Appendix and Heng and others, 2015). - Index components: financial institutions and financial markets; each broken down into access, depth, and efficiency sub-components. - Key observations: - Financial system deepened notably in the past decade driven by growth in financial institutions (improved institutional access and efficiency). Market development stagnated. - Costa Rica lags other emerging markets on all subcomponents of financial market development. - Institutional access outperforms other EM groupings due to a wide network of ATMs and bank branches per 100,000 adults. - Institutional efficiency is low (high interest rate spreads, high overhead costs, high net interest margins) though slightly exceeding the LAC average. - Institutional depth is low: low levels of credit and deposits to GDP and small mutual fund and insurance industries. - Gaps relative to fundamentals: - Costa Rica’s financial development is below levels predicted by its fundamentals on all but two subcomponents (3-bank asset concentration and non-interest income). - Specific inefficiencies: lending-deposit spread, bank interest margin, overhead cost. - Low credit-to-GDP partly reflects a weak legal collateral framework in place before 2015 and a high collateral-to-loan ratio. ### B. Potential for raising growth and stability through further financial deepening - Relationships: - The relationship between financial development and growth, and financial development and stability, is non-linear. - Benefits rise at early stages as resources are channeled to productive uses but there is a turning point after which growth benefits diminish. - At early stages, financial development reduces instability (insurance services), but benefits diminish after a point; excessive borrowing and risk-taking can increase instability and lower long-term growth. - Inverted U-shaped relationship with growth is driven by the depth (size) of financial institutions. - Access and efficiency yield unambiguously increasing benefits to growth, though with potential stability costs as reduced bank profitability may encourage risk-taking. - Too much market development at early stages of institutional development may harm stability (market volatility when institutions weak). - For similar levels of development, institutions and markets are complementary for growth and stability. - Costa Rica-specific position: - Costa Rica has not reached the levels of institutional and market development that yield maximum benefits to growth and stability. - In LAC, Brazil and Chile are closest to reaping those benefits; Dominican Republic, Paraguay, and Honduras lag behind. - Estimates assume other growth determinants are held constant; reaping maximum benefits also requires improving macroeconomic fundamentals. - Causality is difficult to disentangle despite using instrumental variables. ### C. Conclusions and policy recommendations - Summary bullets: - Costa Rica’s financial system deepened notably in the past decade, but continues to lag other emerging markets and the level implied by macroeconomic fundamentals. - Short-term objective: remove distortions preventing full financial development potential given current fundamentals. - Specific policy recommendations: - Modernize secured transactions and collateral framework while balancing with proper regulation and supervision. - Note: In 2015, Costa Rica adopted a new secured transactions law establishing a functional secured transactions system and a modern, centralized, notice-based collateral registry; broadened assets usable as collateral, including intangibles such as intellectual property rights; allowed a general description of assets granted as collateral and permitted out-of-court enforcement. Careful monitoring is warranted as the new system is tested. - Improve efficiency of financial intermediation by ensuring a level playing field for private and public banks. - An important first step: remove the explicit guarantee currently given by the state to all colon-denominated deposits in state banks. - Follow market-friendly debt management and issuance strategies to foster secondary markets for government securities: - Use standardized simple instruments with conventional maturities. - Strengthen legal and regulatory frameworks. - Promote stock market development through: - A more robust macroeconomic environment. - Stronger institutional and legal frameworks to promote investor rights and information disclosure. - Policies that increase market size, in particular supporting development of an institutional investor base. - Recent step: allowing private participation in the insurance sector where it used to be a state monopoly; more entry encouraged. - Strengthen protection of minority investors (Costa Rica scores poorly on Doing Business indicators). - Review tax treatment of securities issuance and investment to make the tax system more attractive to issuers, provided fiscal sustainability is not jeopardized. - Long-term view: as fundamentals evolve (including higher income per capita), further financial development would be advantageous for growth and stability, provided adequate regulatory oversight prevents excesses. Policy should avoid promoting excessive market development when financial institutions are underdeveloped. ### Appendix I. Measuring Financial Development — framework and data - Coverage and data sources: - Data generally cover 1995 to 2013 with gaps; some variables (e.g., ATMs per thousands of adults) available only from 2004. - Sources: World Bank WDI, FinStats, NBFI database, GFD, IMF IFS, Bureau van Dijk Bankscope, Dealogic debt capital markets, WFE, BIS debt securities statistics. - Normalization: - Variables normalized using the formula I_x,it = (x_it - min(x_it)) / (max(x_it) - min(x_it)), and for variables capturing lack of development (interest rate spread, bank asset concentration, overhead costs, net interest margin, non-interest income) one minus the formula above was used. - Weights and aggregation: - Weights estimated by principal component analysis, factor analysis, and equal weights; results robust across methods; index uses equal weights for simplicity. - Regression frameworks: - Regressions use 5-year averages and dynamic panel techniques. - Financial Development Gaps: - Benchmarking regressions link financial development indices (FD, FI, FM) to fundamentals (initial income per capita, government consumption to GDP, inflation, trade openness, educational attainment proxied by average years of secondary schooling for people 25+, population growth, capital account openness, size of the shadow economy, rule of law). - Instruments Z_it used: rule of law and legal origin dummies. - Predicted norms computed: FI_it = δ’1 X_it^FI + δ’2 Z_it + h^FI_t + e^FI_it. Gaps = actual index minus calculated norms. - Financial development, growth, and stability: - Real GDP growth (ΔY_it) linked to financial development allowing non-linearity by adding square of financial development; controls same as benchmarking plus ratio of FDI to GDP and capital account openness. - Financial instability (FS_it) measured by first principal component of inverse distance to distress (z-score), real credit growth volatility, and real and nominal interest rate volatility. - Growth volatility (GV_it) measured by standard deviation of GDP growth. - Controls for instability regressions: initial income per capita, government consumption to GDP, trade openness, changes in terms of trade, growth in per capita income, capital flows to GDP, exchange rate regime, political stability, indicator for offshore financial center. - Estimation methods: - System GMM estimation (Arellano and Bover, 1995; Blundell and Bond, 1998) used to address dynamic dependence and potential endogeneity. Instruments used include rule of law (Kaufmann, Kraay and Mastruzzi 2010) and legal origin dummies (La Porta, Lopez-de-Silanes and Shleifer 2008). - The following functional forms estimated: - Aggregated: f(FD_it) = b1 FD_it + b2 FD_it^2 - Subcomponents: f(FI_it, FM_it) = b1 FI_it + b2 FI_it^2 + b3 FM_it + b4 FM_it^2 + b5 FI_it x FM_it - Estimated equations: table with coefficients (excerpt) - Dependent variables: Financial Instability; Growth Volatility; Growth - Sample coefficients (selected): - FD: -6.457* (3.814) for Financial Instability; -21.42*** (7.270) for Growth Volatility; 11.47* (6.279) for Growth - FD2: 6.263 (5.735); 23.74** (10.82); -12.38* (6.556) - ΔFD: 5.283** (2.160); 8.423** (4.008); 5.698* (3.075) - FI: -13.75** (5.419); -27.89*** (9.533); 30.83*** (8.788) - FI2: 18.64** (8.123); 36.38** (14.45); -48.36*** (11.58) - FM: -0.772 (3.119); -6.779 (5.345); -0.586 (3.987) - FM2: 3.360 (4.886); 18.02** (8.324); -12.35** (5.314) - FM*FI: -5.140 (9.730); -5.354 (15.81); 27.27** (13.16) - ΔFI: 4.753** (2.114); 14.08*** (3.708); 7.088** (2.958) - ΔFM: 3.190* (1.672); -2.335 (2.846); 0.508 (2.222) - Observations: 143; 143; 158; 158; 301; 301 (as reported in table) - Note: Standard errors in parentheses; significance levels indicated: *** p<0.01, ** p<0.05, * p<0.1 - Variables included in the Financial Development Index (Institutions and Markets; Access, Efficiency, Depth): - Bank branches per 100,000 adults - ATMs per 100,000 adults - Domestic Bank Deposits / GDP (%) - Insurance Company Assets / GDP (%) - Mutual Fund Assets / GDP (%) - Domestic Credit to Private Sector / GDP (%) - Bank concentration (%) - Bank lending-deposit spread - Overhead cost/total assets - Bank net interest Margin - Non-interest income/total income - Total number of issuers of debt (domestic and external, NFCs and Financial) - Market capitalization excluding top 10 companies to total market capitalization - Stock market capitalization to GDP - Stock market total value traded to GDP (%) - Stock of government debt securities in % of GDP - Debt securities of financial sector by local firms in % of GDP - Debt securities of non-financial sector by local firms in % of GDP - Stock market turnover ratio (value traded/stock market capitalization) *International Monetary Fund — Costa Rica: References and accompanying Financial Deepening, Financial Inclusion, and Appendix material contained in the supplied content unit.* ### 1. The benefits of financial inclusion could be notable in Costa Rica. Financial inclusion ### 1. The benefits of financial inclusion could be notable in Costa Rica. ### Overview - Financial inclusion can boost economic growth and reduce poverty and inequality by mobilizing savings and providing households and firms with greater access to resources needed to finance consumption and investment and to insure against shocks. - Financial inclusion can foster labor and firm formalization, help boost government revenues, and strengthen social safety nets. - Given modest potential growth, rising inequality, and low savings and investment in Costa Rica, the benefits from further financial inclusion could be pronounced. - Note preparation: Prepared by Anna Ivanova, Yixi Deng, and Joyce Wong. ### Empirical findings: household vs. enterprise inclusion - Costa Rica has excelled on financial inclusion of households but lags behind peers on inclusion of enterprises and access to financial infrastructure. - Recent household improvements: largest gains in account holdings, usage of ATMs, and debit cards. - Costa Rica is one of the champions on household financial inclusion among LAC countries and outperforms peers on all subcomponents of the household financial inclusion index. - Enterprise inclusion shortcomings: - High collateral requirements compared to other emerging markets. - Access to and cost of finance seen as a major constraint by a large share of SMEs. - Relatively weak access to financial institutions as measured by number of bank branches and ATMs in relation to population and area. - Mobile phone–based solutions and correspondent banking measures also compare poorly to peers. ### Regulatory environment and the Global Microscope - Costa Rica scores below the average of other emerging markets on the Global Microscope index (assesses regulatory environment for financial inclusion across 12 indicators and 55 countries). - Relative strengths: prudential regulation and credit reporting systems. - Areas of underperformance (all other Microscope indicators): - regulation and supervision of deposit-taking activities, - regulation and supervision of branches and agents, - regulation and supervision of credit portfolios, - grievance redress and operation of dispute resolution mechanisms, - market conduct rules, - requirements for non-regulated lenders, - regulatory and supervisory capacity for financial inclusion, - regulation of electronic payments, - regulation of insurance for low-income populations, - government support for financial inclusion. - The Microscope score appears more relevant for explaining inclusion of enterprises than household inclusion. ### Financial inclusion gaps relative to fundamentals - Method: Gaps computed as deviations of financial inclusion indices from values predicted by exogenous domestic factors (income per capita, education, size of the shadow economy, rule of law, share of foreign-owned firms, importance of fuel exports). - Interpretation: Negative gaps may capture distortions or market frictions; positive gaps may reflect financial excesses. - Determinants associated with higher financial inclusion (from Dabla‑Norris et al, 2015): - higher income per capita (for households and firms), - higher education (for households), - stronger rule of law (for households), - lower degree of informality (for households), - lower prevalence of foreign-owned firms (for firms and access to financial institutions), - lower fuel exports (for firms and access to financial institutions). - Costa Rica’s gaps: - Positive financial inclusion gap for households overall—positive gaps on ATM and debit card usage outweigh negative gaps on savings and borrowing from a financial institution and usage of credit cards. - Positive household gap relative to an Asian benchmark as well; no obvious indications that household positive gap reflects excesses or inefficiencies. - Negative financial inclusion gap for enterprises overall—driven by high collateral, high share of firms identifying access to/cost of finance as constraint, and low usage of banks by SMEs for investment and working capital; positive gaps on savings and checking account holdings and high share of SMEs with a loan or credit line are outweighed. - Large negative gap on collateral required for a loan linked to weak legal collateral framework. - High cost of financing partly due to substantial presence of public banks, which lack incentives to improve efficiency. - Negative gap relative to an Asian benchmark on the firm side as well. ### Econometric evidence on drivers of gaps and policy implications - Regression results indicate higher (more positive/less negative) financial inclusion gaps are associated with: - lower non-interest income (for household and firms), - lower bank safety buffers (for households), - lower bank efficiency, as measured by overhead costs (for firms), - stronger regulatory environment, as measured by the Global Microscope score (for firms). - Causality is not certain—e.g., bank safety buffers and efficiency could be results of increased inclusion rather than drivers. - Policy implication emphasized for Costa Rica: strengthening the regulatory environment for financial inclusion (Global Microscope dimensions) could help improve inclusion of firms. - Specific recommendation: put in place regulatory incentives to formalize micro-lenders/savers (currently constituted as non-regulated NGOs) and adopt explicit regulations related to banking agents, e-money, and micro-insurance to bring existing activities into the formal regulatory perimeter. - Analysis is partial; other factors are likely at play. ### Theoretical framework and policy experiments - A micro-founded structural model is applied to examine the implications of relaxing financial constraints for growth and inequality. - Financial constraints grouped into three dimensions: - Participation costs: high documentation requirements and barriers (red tape, informal guarantors) that impede access to finance. - Borrowing constraints: depth of credit determined by collateral requirements, influenced by creditors’ rights, information disclosure, and contract enforcement. - Intermediation costs: high costs from information asymmetries and limited competition leading to higher interest rates and fees for smaller/less capitalized borrowers. - Model calibration targets: percent of firms with credit, firm employment distribution, economy-wide nonperforming loan ratio, and interest rate spread. - Three illustrative simulations: - (i) reducing the financial participation cost to 0, - (ii) relaxing borrowing constraints in the form of collateral requirements to the world minimum, - (iii) increasing intermediation efficiency (reducing monitoring costs to 0 by equalizing spreads to the proportion of non-performing loans). ### Model results: growth–inequality tradeoffs and country-specific implications - General effects of lowering financial constraints: - GDP increases via: more firms accessing formal banking, more capital invested, more efficient allocation of funds, higher productivity as best firms scale up, and reduced waste on contract negotiation. - Variation in impact: depends on country-specific characteristics and which constraint is alleviated. - Highest growth dividends from relaxing collateral requirements, but this may increase inequality because most benefits accrue to already productive entrepreneurs with existing credit access and higher incomes. - Inequality effects could be mitigated if relaxation benefits previously constrained entrepreneurs. - Reducing intermediation costs produces similar trade-offs to relaxing collateral requirements. - Reducing participation costs tends to benefit those outside the financial system more and thus decrease inequality. - Participation cost is a fixed cost and a relatively lower share of income for already-included entrepreneurs, so it benefits new entrepreneurs more. - Overall welfare: the “poor” may still be better off under lower borrowing constraints, though with smaller gains than the “rich.” ### Key statistics and comparative figures (preserved exactly) - Regression explanatory power: R-squares close to 0.7 in regressions for households and firms. - Three illustrative simulations include: (i) reducing the financial participation cost to 0, (ii) relaxing borrowing constraints in the form of collateral requirements to the world minimum (iii) increasing intermediation efficiency (i.e., reducing monitoring costs to 0 by equalizing spreads to the proportion of non-performing loans). - Share of firms with access to credit in Costa Rica: close to 60 percent. - Collateral needed for loans by firms in Costa Rica is 1.5 times higher in Guatemala, for example. - Interest rate spreads in Costa Rica are 2.5 times higher in Costa Rica than in Panama and El Salvador. *Prepared by Anna Ivanova, Yixi Deng, and Joyce Wong.* ### 11. Costa Rica could raise growth by relaxing collateral constraints and lowering ### _cr16132 - 11. Costa Rica could raise growth by relaxing collateral constraints and lowering ### Key findings - Collateral levels in Costa Rica are so high that a significant number of firms remain constrained in the amount of borrowing they can access (even those which have a positive level of credit). - Relaxing collateral constraints would: - Raise growth. - Lower inequality, because benefits accrue not only to top firms but also to a large mass of medium-sized firms that expand, driving up workers’ wages. - Reducing monitoring/intermediation costs would: - Provide high growth benefits. - May lead to a moderate increase in inequality, since benefits accrue mainly to firms that already have credit and tend to be higher-income entrepreneurs. - The differing effects on inequality arise from effects on the extensive and intensive margins of credit: - Lowering collateral constraints affects both extensive (bringing new entrepreneurs into credit) and intensive margins (relieving constraints on some existing borrowers). - Reducing intermediation costs affects mainly the intensive margin, benefiting already-credit-accessible entrepreneurs. ### Model and mechanisms - Model features: - Heterogeneous agents differing in talent and wealth. - Agents choose between being workers (earn wages) or entrepreneurs (earn profits) and whether to pay a fixed participation cost to access banking. - Participation cost represents the cost of financial contracting; higher participation cost increases credit autarky and disproportionately excludes poor but talented individuals. - Entrepreneurs choose investment, borrowing, and employment; output depends on capital, labor, and talent. - Two key frictions once in the banking system: - Borrowing constrained by collateral due to imperfect enforceability of contracts; weak contract enforceability leads to lower leverage. - Asymmetric information leading to monitoring/audit costs; banks recuperate these through interest rates and high overhead fees. - Calibration and data: - Baseline calibrated to data for 12 LAC countries. - Firm-level data for 2005 from the World Bank Enterprise Survey, plus macro and financial variables (savings rate, non-performing loans (NPLs), interest rate spreads) for 2010 or latest available. - Model focuses primarily on formal sector firms. - Key parameters jointly chosen to match simulated moments (percent of firms with credit, firm employment distribution). ### Quantitative impacts (model simulation results highlighted in figures) - Effects illustrated in Figure 5 (authors’ calculations): - Effect on GDP of making participation costs 0: shown as % pts change in GDP for various countries including CRI. - Effect on GDP of taking collateral to world minimum: shown as % pts change in GDP for various countries including CRI. - Effect on GDP of reducing monitoring costs to 0 (making spreads = NPL): shown as % pts change in GDP. - Effect on Gini of making participation costs 0: absolute change in GINI for various countries including CRI. - Effect on Gini of taking collateral to world minimum: absolute change in GINI. - Effect on Gini of reducing monitoring costs to 0 (making spreads = NPL): absolute change in GINI. - Figures show that for Costa Rica: - Reducing collateral to world minimum and eliminating participation costs generate positive % pts changes in GDP and reductions in GINI. - Reducing monitoring costs increases GDP but can raise GINI modestly (exact plotted values are in the figures). ### Determinants and empirical evidence - Costa Rica’s performance: - Has excelled in financial inclusion of households but lags peers on inclusion of enterprises and access to financial physical infrastructure. - Made important strides in household inclusion with largest improvements in account holdings, usage of ATMs, and debit cards. - Enabling environment and constraints: - Performs well on prudential regulation and is ahead of other LAC countries on credit reporting systems. - Large portions of the financial system remain outside the regulatory perimeter, hindering further inclusion. - Positive financial inclusion gap for households but negative gap for enterprises after accounting for macro fundamentals. - Causes of the negative firm inclusion gap: - High value of collateral required for a loan linked to weak legal collateral framework. - High cost of financing related to: - Dominance of public banks that lack incentives to improve efficiency. - Uneven playing field for private banks that stifles competition. - Weak regulatory environment for financial inclusion. ### Empirical regressions and statistics (selected reported results) - From regression tables (examples of coefficients and significance): - log_GDP_pcap: coefficients 0.0918***, 0.104***, 0.111***, 0.142***, 0.131*** (standard errors in parentheses). - Mean years of schooling (of adults): 0.0173**, 0.0160***, 0.0162**, 0.00927, -0.0189. - Rule of Law (-2.5(weak) to 2.5(strong)): 0.0733***, 0.0657***, 0.0502**, 0.0670, 0.0330. - Observations: 781, 588, 045, 111 (as reported in table rows). - R-squared values reported: 0.732, 0.708, 0.738, 0.658, 0.215. - Table on financial sector determinants (selected coefficients): - Non-Interest Income / Total income (%): -0.00428**, -0.00515*, 0.00272. - Bank net interest margin (%): -0.0144, -0.0202, 0.0399. - Overhead Costs / Total Assets (%): 0.0183, 0.0320*, -0.0216. - Distance to default: -0.00261**, -0.00243, -0.000316. - Observations: 433, 046, 0 (as reported) with R-squared: 0.200, 0.268, 0.154. - Appendix tables and figures show regional indices (Household Inclusion Index, Firm Inclusion Index, Access Index) with region-level numeric totals (e.g., Household Inclusion Index Total 104 for 2011 and 2014; Firm Inclusion Index Total 79; Access Index Total 161). ### Policy recommendations - Primary focus: - Further modernize the collateral framework. - Strengthen the regulatory environment for financial inclusion. - Complementary actions: - Ensure a level playing field for private banks compared to public banks to help lower intermediation costs; this could provide additional growth benefits but carries the potential for increased inequality. - General insight: - Micro-founded general equilibrium model suggests a potential trade-off between the objectives of growth and inequality reduction: relaxing collateral constraints raises growth and reduces inequality; lowering intermediation costs raises growth but may increase inequality. ### Data, indices, and methodology notes - Financial Inclusion definition used: access and effective usage of financial services by households and firms. - Constructed three multi-dimensional indices: - Usage of financial services by households (Findex). - Usage of financial services by SMEs (Enterprise Survey). - Access to financial institutions (Financial Access Survey). - Normalization formula for variables: Ix,it = (xit − min(xit)) / (max(xit) − min(xit)); for variables capturing lack of inclusion (e.g., value of collateral) the reverse formula Ix,it = 1 − (xit − min(xit)) / (max(xit) − min(xit)) was used. - Weighting and aggregation: - Explored principal component analysis, factor analysis, and equal weights. - For simplicity and robustness, results presented with equal weights. - Data sources include Global Findex (2011 and 2014), World Bank Enterprise Survey (latest observation), IMF Financial Access Survey, and WDI for physical infrastructure measures. *Source: IMF staff calculations and text in content unit _cr16132 - 11. Costa Rica could raise growth by relaxing collateral constraints and lowering* ### 1. Fiscal sustainability remains elusive as debt continued its rapid ascent, with the ### 1. Fiscal sustainability remains elusive as debt continued its rapid ascent, with the primary fiscal deficit stabilizing at post-crisis peak levels. ### Recent fiscal developments - Central government fiscal position deteriorated sharply during the global credit crisis: - Primary and overall fiscal deficits reached 3 and 5½ percent of GDP by 2010, respectively. - Deterioration driven by endogenous fall in revenue after above-trend GDP growth pre-2008-09 and sharp increase in expenditure (mainly wages and transfers) from countercyclical policies. - Fiscal restraint efforts in 2011 (mainly cuts in capital expenditure) were undermined by rising transfers and interest bill in 2012-13. - A tax reform aimed at sustainability was voided by the Supreme Court in 2012; revenues stagnated. - Central government primary and overall deficits returned to post-crisis peaks by 2013. - New administration (mid-2014) pursued broadly neutral fiscal policy: - Primary deficit broadly flat at 3 percent of GDP in 2014, aided by substantial expenditure under-execution and strong efforts to reduce tax evasion in 2015. - Public debt and financing costs: - Public debt rose to 42½ percent of GDP in 2015, up from 24 percent of GDP in 2008. - Interest bill increased only moderately, keeping overall deficit around 6 percent of GDP in recent years—enabled by relatively low financing costs and historically low international rates. - In 2015, domestic government financing costs declined further, especially at the short-end of the yield curve, amid strong disinflation and aggressive monetary policy response. - Other government levels: - CCSS (social security) maintained a small surplus of 0.75 percent of GDP in 2015. - Public sector enterprises turned to a small deficit of 0.1 percent of GDP (mainly refinery losses). - Central bank operational deficit about 0.75 percent of GDP. - Analytical focus: - Fiscal vulnerabilities, medium-term sustainability issues, and related adjustment needs are analyzed at the central government level (while briefly discussing CCSS longer-term sustainability). ### Sustainability gap and recommended adjustment - Required adjustment to restore debt sustainability: - A total adjustment of about 3¾ percent of GDP is necessary to restore debt sustainability. - In a passive scenario, primary deficit projected to rise to 3¼ percent of GDP in 2016, and to 3½ over the medium term as the constitutionally-mandated education expenditure target of 8 percent of GDP is gradually achieved. - Staff analysis indicates a debt-stabilizing primary surplus of ¼ percent of GDP is needed to stabilize debt in the medium term given projected increase in real rates above real GDP growth. - Timing matters: - Further postponing fiscal retrenchment increases the primary balance improvement required to stabilize the public debt ratio. - Optimal consolidation path: - A gradual but frontloaded fiscal consolidation is recommended to balance closing the sustainability gap and limiting adverse growth impact. - Compared with the 2014 Article IV model results, the optimal path is somewhat less frontloaded due to a moderate widening of the negative output gap in 2015-16; now a bit less than one third of the adjustment is recommended in the first year. - Staff report’s adjustment scenario spreads the adjustment more into the second and third years and assumes the full sustainability gap closed by end of the 5-year projection period. - Key fiscal parameters and gap (Fund staff estimates): - Medium-term primary balance (I): -3.5 - 2015 outturn: -3.0 - 2016, projected change: -0.2 - 2017-21, projected change: -0.2 - Debt stabilizing primary balance (II): 0.3 - 2021, real growth (in percent): 4.0 - 2021, real interest rate (in percent): 4.6 - 2021, debt: 55.5 - Fiscal Sustainability Gap (II - I): 3.8 - Long-term scenario on primary balance, overall balance, and central government debt (presented as baseline, gradual adjustment, immediate adjustment paths in staff analysis): - Gradual scenario assumptions: authorities place 90 percent weight on growth objective; fiscal multiplier 0.3; self-correction parameter for output gap 0.5. - Immediate adjustment scenario assumption: full fiscal adjustment in 2016 with no impact on growth. - Pension actuarial deficit implication: - Including the actuarial deficit of the public pension system raises required fiscal adjustment by an additional 1½ percent of GDP. - CCSS currently runs a cash surplus of about ¾ percent of GDP in 2015 but projected to turn cash deficit medium/long term due to maturation and aging. - Simulations indicate reforms equivalent to about 1½ percent of GDP required to achieve actuarial balance over next 50 years; actuarial gap increases to almost 4 percent of GDP over 100 years. - Reform options: higher contributions, reduced replacement rates, and/or increase in retirement age. - Notes on CCSS coverage and legacy regimes: - IVM covers ⅔ of the labor force (approximately 1½ million workers) and has about 165,000 beneficiaries. - Legacy regimes for civil servants and teachers are financed with central government transfers reaching about 2½ percent of GDP in 2015. ### Policy mix and composition of consolidation - Consolidation requires action on both revenue and expenditure sides: - Revenue mobilization should be cornerstone given generally low tax revenues compared to other middle-income countries. - Authorities’ plans and implementation status: - Previous administration prepared a consolidation plan of about 3½ percent of GDP; not implemented before 2014 elections. - New administration developed a plan broadly in line with staff recommendations on size, composition, and pace, though expenditure measures need further clarification. - Measures submitted to Parliament (authorities’ proposals and staff assessment): - Government submitted measures expected to yield some 2¾ percent of GDP: - Revenue measures (~slightly more than 2 percent of GDP): VAT and income tax reforms; other provisions (corporate income tax amendments and anti-tax evasion measures) generating almost ½ percent of GDP. - VAT reform: broaden base to include services; gradual increase in rate from 13 to 15 percent starting in 2016; separate increases in taxes on sales of vehicles and real estate; radical reduction in basic goods basket conditional on transfers system to be revenue-neutral for lower-income households. - Income tax reform: two additional marginal rates of 20 and 25 percent on higher-income brackets; unify taxation of income on capital at 15 percent; introduce taxation of capital gains. - Expenditure provisions presented to Congress would reduce outlays by ¼ percent of GDP in 2016 (caps on budgetary pensions; paring down transfers to decentralized institutions). - Government agreement with main opposition to make miscellaneous cuts of ¼ percent of GDP via supplementary 2016 budget. - Authorities identify administratively-determined spending cuts to contain current spending growth (mostly transfers and public sector wages) to keep it through medium term below expansion of nominal GDP; intend this to yield remaining adjustment. - Authorities presented a fiscal rule proposal aimed at preserving government debt sustainability; broadly aligned with two-pillar frameworks (anchor e.g., debt; operational target e.g., expenditure) but key elements require greater specification. - Scenarios and outcomes: - Baseline scenario (measures with higher probability of approval + agreed expenditure cuts) implies partial adjustment of about 2¼ percent of GDP: - Under this, central government fiscal deficit would decline moderately to around 5½ percent of GDP. - Public debt ratio would continue to increase to almost 55 percent of GDP by 2021. - Full adjustment scenario (complete fiscal adjustment to restore debt sustainability): - A tighter fiscal stance restoring debt sustainability could be achieved without significantly affecting growth, allowing for a more balanced macro policy mix with looser monetary policy consistent with the inflation target. - Emphasis on revenue measures that increase tax progressivity and target higher earners (higher income tax rates on higher brackets, higher VAT with transfers to protect low-income households, anti-tax evasion) is likely to limit negative growth impact. - Frontloaded consolidation would mitigate increases in market rates associated with US monetary policy normalization and reduce the current account deficit. ### Key quantified consolidation matrix (authorities’ plans, staff assessment, and additional measures; in percent of GDP) - Total adjustment (Total): 4.0 (Authorities' plans) | 2.9 (Baseline Scenario (I)) | 2.3 (Additional Measures (II)) | 1.5 (Total adjustment (I + II)) | 3.8 (Total adjustment) - Revenue: 2.7 | 2.5 | 1.7 | 0.8 | 2.5 - Anti-tax evasion: 0.5 | 0.2 | 0.2 | -0.2 - VAT: 1.3 | 1.3 | 0.5 | 0.8 | 1.3 - Income tax: 0.6 | 0.6 | 0.6 | -0.6 - Corporate income tax: 0.2 | 0.2 | 0.2 | -0.2 - Vehicles and real estate tax: 0.2 | 0.2 | 0.2 | -0.2 - Expenditure: 1.3 | 0.4 | 0.6 | 0.7 | 1.3 - Transfers: 0.4 | 0.4 | 0.4 | -0.4 - Wages: 0.6 | - | - | 0.6 - Miscellaneous cuts in 2016 budget: 0.2 | - | 0.2 | -0.2 - Goods and services: 0.1 | - | 0.1 | 0.1 - Notes on table: - Staff assessment on revenue includes only measures deemed to have higher probability of approval; baseline assumes VAT rate increases will not be approved by Congress. - VAT reform in full adjustment envisages full move from sales tax to VAT, extend coverage to services, and gradual increase from 13 to 15 percent. - Income tax reform details as above. - Corporate income tax amendments aimed at reversing budgeted decline due to Constitutional Court ruling on 2010 amendments. - Transfers include cuts to transfers for capital expenditures of decentralized institutions and enforcement of legal cap on pensions paid out of the budget. - Wages measures include containing nominal growth of public wages, freeze in hiring outside education, and cuts to public compensation bonus schemes. - Miscellaneous cuts reflect agreement between ruling party and main opposition for supplementary 2016 budget. ### Risks and mitigating factors - Political agreement is critical: - Scenarios of partial and full implementation underscore the importance of reaching political agreement on fiscal reforms to close the sustainability gap. - Advantages of full implementation: - Full adjustment could permit a more balanced macro policy mix, lower market-rate pressure from US monetary policy normalization, and reduction in current account deficit. - Distributional and growth considerations: - Emphasis on revenue measures that increase progressivity and target higher earners is projected to limit adverse consumption effects and growth impact. - Frontloaded adjustment preferred to mitigate financing cost pressures. *Prepared by Jaume Puig-Forné. Sources: National authorities and Fund staff estimates.* ### 9. There are substantial upside risks to the ### 9. There are substantial upside risks to the ### Debt dynamics and shock sensitivities - A plausible macro-fiscal shock as defined in the Fund’s DSA framework for market access countries would result in central government debt rising above 65 percent of GDP by 2021, more than 10 percent above the level in the baseline scenario (DSA Annex to the staff report). - Debt dynamics are most sensitive to a growth shock: - An isolated one standard deviation shock to growth in 2016-17 results in an increase in central government debt of about 5 percent of GDP by 2021 relative to the baseline scenario. - Fiscal shock sensitivity: - A fiscal shock equivalent to an additional 1¼ of GDP increase in the primary deficit in 2016–17 would increase debt by about 3 percent of GDP by 2021. - Interest rate sensitivity: - A 200 basis points shock to the average real interest rate raises the debt-to-GDP ratio by less than 2 percentage points of GDP by 2021. - Exchange rate sensitivity: - A 15 percent depreciation in the nominal exchange rate has an impact on debt of less than 1 percent of GDP by 2021. - Combined macroeconomic shock definition (as used in the DSA): - Fiscal shock: equivalent to 50 percent of planned cumulative adjustment or to half of a standard deviation of historical observations of the primary balance, whichever is greater. - Growth shock: 1 standard deviation shock to real GDP growth for 2 consecutive years. - Interest rate shock: nominal interest rate increase by the difference between the maximum real interest rate over the last 10 years and the average real interest rate over the projection period, or a 200 basis point shock, whichever is larger. - Exchange rate shock: correction of the Fund’s estimate of real exchange rate overvaluation, or maximum historical depreciation of the exchange rate, whichever is the highest. ### Sovereign financing risks and market access - Gross financing needs and access risks: - Average projected financing needs of close to 10 percent in 2015-16, among the highest of CAPDR and LA-5 countries. - Market indicators and ratings: - Spreads on external sovereign bonds have increased substantially since pre-US tapering troughs and are now among the highest of CAPDR and LA-5 countries. - Costa Rica lost its only investment grade rating in 2014. - After the recent loss of its only investment grade rating by Moody’s, the sovereign is now rated one (Moody’s and Fitch) to two notches (S&P) below investment grade. - Upside risks to funding costs: - Uncertainties about Congressional approval of new external bond issuance and expected upward normalization of global interest rates suggest financing costs are likely to increase. - The risk of a debt spiral with potentially non-linear increases in financing costs cannot be discarded if debt continues to rise unabated. ### Domestic investor base, bank exposures, and financial stability - Evolution of domestic holdings: - In 2008, domestic institutional investors (including the CCSS, nonfinancial public sector institutions, and banks) held about 80 percent of domestic government bonds. - By 2014, this share had fallen to 60 percent. - The non-financial private sector disproportionately absorbed the increase in debt resulting from the worsening fiscal position. - Banking sector exposure and stress scenarios: - Simulations of a similar increase in domestic sovereign yields as in 2012 show that banks could be close to failing regulatory capital requirements in the medium term from mark-to-market losses on their growing sovereign exposures in a passive scenario of no fiscal consolidation. - Losses would be even larger under more plausible assumptions of much higher government financing costs under this scenario. - Potential bank capital losses illustration: - Charted scenario: Loss from 400 bps increase in domestic bond yields; regulatory capital after loss shown against regulatory capital requirement (values in percent of risk weighted assets noted in source figures). ### Balance sheet analysis (BSA) — external and sectoral positions - Net external positions (selected points): - Total net external debtor position of about 42 percent of GDP in 2014, up from 35 percent in 2013. - Excluding FDI liabilities, the economy has a small net creditor position of about 3 percent of GDP in 2014, down from 4 percent in 2013 and 8 percent in 2010. - Drivers of changes: - The 2014 decline in the net creditor position excluding FDI was driven mostly by continued external debt issuance by the central government. - Increased reliance on external financing by the financial sector was an important driver of the decline from 2010 to 2013. - Currency mismatch and FX exposures: - Risks from currency mismatches appear limited at the aggregate sectoral level. - Unhedged FX borrowers in the private sector remain a key risk. - The financial sector remained a net foreign currency creditor—creditor position of 1½ percent of GDP excluding FDI—while channeling additional external financing into domestic credit in FX. - Household and private sector leverage trends: - Bank credit to the private sector increased from less than 30 percent of GDP in 2004 to almost 55 percent in 2015. - Credit to non-financial corporates increased from 11 to 17 percent of GDP (2004 to 2015). - Credit to households increased from 20 to 35 percent of GDP (2004 to 2015). - The bulk of the increase in credit to households has been in domestic currency. - The share of credit in FX to households fell appreciably in 2004-09 and remained fairly stable around 30 percent since then. - The share of credit in FX to the corporate sector is currently about 65 percent of total credit to corporate. - Table 1 excerpts (End-2014, in percent of GDP): - Economy: Gross External Assets 38.0; Gross External Liabilities 79.9; Net External Position -41.9; Net External Debt Position 2/ 2.8. - Net FC Position (economy): -42.1. - Financial sector (End-2014): Net External Position -8.9; Net External Debt Position 1/ -6.6. - Sources: Banco Central de Costa Rica, and Fund staff estimates. - Note: 1/ Excluding net FDI position. ### Selected real sector issues — potential output - Potential output and output gap estimates (selected values): - Staff analysis suggests potential output growth in 2015 is about 4.0 percent and the output gap somewhat negative. - For the period 1999–2008, potential output grew at an average rate of 4.6 percent. - Post-crisis (2009-14) estimate: 3.9 percent average (recovering moderately). - Output gap estimated to have broadly closed in 2013 (-0.3 percent) and widened in 2015 to -1.1 percent. - Range across methods for 2015 potential growth: - Cycle extraction filters and univariate Kalman filters suggest potential output growth is between 3.8 and 4.6 percent. - Multivariate filter including inflation and financial variables suggests potential output growth about 3.6 percent. - Production function approach indicates main drivers of potential GDP growth are capital and labor supply; potential output grew at an average rate of 4.8 percent in 2000-2008. - Table 1 summary (selected rows, potential GDP growth rate and output gap): - Production Function: 2015 potential growth 3.9; 2014 output gap 0.25; 2015 output gap -0.36. - Hodrick-Prescott: 2015 potential growth 3.8; 2014 output gap 0.46; 2015 output gap -0.06. - Univariate Kalman (Mean Reversion): 2015 potential growth 4.6; 2014 output gap -0.40; 2015 output gap -1.15. - Multivariate Kalman (With inflation): 2015 potential growth 3.64; 2014 output gap -0.02; 2015 output gap -0.34. - Average of All Models: 2015 potential growth 3.97; 2014 output gap 0.22; 2015 output gap -0.35. - Macroframework: 2015 potential growth 4.00; 2014 output gap -0.43; 2015 output gap -1.11. - Specific shock from Intel exit: - Intel’s withdrawal lowered actual and potential real GDP by about ¾ percentage points over 2014-2015. - Level effect on potential output of Intel exit estimated at 0.2 percent of GDP in 2014 and 0.8 percent in 2015. *Source: IMF staff report excerpts (Costa Rica).* ### 5. In general, the estimates of TFP growth should be interpreted with caution. The TFP ### 5. In general, the estimates of TFP growth should be interpreted with caution. ### A. TFP measurement: caveats and proximate causes of low productivity - TFP is defined as a residual—the difference between output growth and the growth in the quantity (and quality) of inputs; therefore, measurement errors in labor and capital series are automatically attributed to TFP. - Factors that can distort measured TFP include: - migration of skilled labor; - mismeasured changes in the quality of the capital and labor stocks; - changes in the level of capital utilization; - use of land. - Indicators consistent with genuinely lower productivity growth in Costa Rica: - Informality: since 2011 the number of people employed in the informal sector has increased by 40 percent since 2011, while the number of workers in the formal sector has not changed. - R&D spending: the economy as a whole spends less than ½ percent of GDP on R&D; by comparison the OECD average is around 2.5 percent of GDP and some emerging economies spend more than 4 percent of GDP. - Education and infrastructure quality/efficiency: - Costa Rica spends around 7-8 percent of GDP on education, but has education outcomes similar to Chile, which spends almost 50 percent less on education as percentage of GDP. - According to the World Economic Forum survey, Costa Rica’s quality of infrastructure is one of the worst among OECD countries. ### B. Impact of Intel exit (2014–15) - Authorities’ estimates: - Total effect of Intel’s withdrawal was about ¾ percent of GDP over 2014-15. - Shutdown lowered real and nominal output by 0.2 percent of GDP in 2014 and a further 0.6 percent in 2015. - Estimates use recently adopted chained real GDP data with 2012 as the reference year. - Outturns are in line with staff projections from the 2014 Article IV staff report. - Staff assessment on potential output and longer-term effects: - Staff estimates the effect of the exit on potential output in 2014-15 is similar to the effect on actual output (¾ percent) with virtually no effect on the output gap. - Loss of manufacturing operations reduces possible future productivity spillovers from Intel to the rest of the economy. - Countervailing prospects: other technology and medical device manufacturing companies maintain or strengthen presence; Intel retained and expanded R&D operations in Costa Rica, hiring additional R&D personnel after manufacturing departure. ### C. Real sector spillovers: sensitivity to external demand shocks - Multi-country VAR model results (assessing sensitivity of Costa Rica GDP growth to trading partner shocks, accounting for cross-country spillovers): - A 1 percentage point reduction in domestic demand growth in the U.S. over a year leads to a maximum reduction in growth in Costa Rica by over 1 percentage points. - A 1 percentage point reduction in growth in other Central American countries over a year would lower Costa Rica growth by about ¾ percentage points. - A 1 percentage point reduction in Chinese domestic demand growth lowers Costa Rica growth by less than ½ percentage points. - A similar reduction in Brazilian domestic demand growth barely has any effect on Costa Rica growth. - Implication: Costa Rica is more vulnerable to growth trends in the U.S. and Central America than to shocks originating in large emerging markets such as China and Brazil. ### Box 1. Methodologies for potential output estimates (key specifications and parameters) - Production-function approach: - Cobb-Douglas: Yt = At Kt^α Lt^(1−α) - α is set at 0.35 (output elasticities sum to one). - Labor force data up to 2010 from Penn World Table 7.1 (PWT); assumed to grow at the 2000-10 average thereafter. - Capital stock: perpetual inventory method Kt = (1−δ) Kt−1 + It with δ = 0.05. - Initial capital stock K0 = I*/(g+δ), where I* is benchmark investment (average share of investment in GDP) and g is average economic growth over 1991-2013. - TFP residual: At = Yt/(Kt^α Lt^(1−α)). - Univariate filters: - HP filter restriction parameter set at 6.25 for annual data (equivalent to 1600 for quarterly data). - Two univariate Kalman filter specifications: - First: potential output follows a random walk with deterministic drift μ̅; output gap modeled as AR(2). - Second: allows mean reversion in the drift with adjustment coefficient β ∈ (0,1); μt = (1−β) μ̅ + β μt−1. - Multivariate filter (augmented with inflation): - Follows Laxton et al. (2010) specification; decomposes log(GDPt) into a sustainable level and cyclical component and is augmented by a Phillips curve πt = λ πt+1 + (1−λ) πt−1 + β yt + επt. - Incorporates inflation and growth expectations (πt+j^C, Gt+j^C) with estimation by Bayesian maximum likelihood using informative priors. - This method estimates a level of “sustainable” output defined as the level of potential output that does not trigger inflation (Okun’s definition), directly pertinent for inflation-targeting monetary policy. *Source: _cr16132 - 5. In general, the estimates of TFP growth should be interpreted with caution.* ### 11. Cross-country regression results are consistent with many of the findings in the micro- ### 11. Cross-country regression results are consistent with many of the findings in the micro- ### Cross-country regression findings - Cross-country regressions support the importance of: - Investment in infrastructure (transportation and telecommunications). - Presence of children proxied by higher fertility rates. - Higher education levels (especially female tertiary enrollment). - Internet access. - Investments in transportation and telecommunications have positive and significant coefficients; internet access coefficients are positive and are stronger when female LFP under the age of 25 is considered. - Female tertiary enrollment share has positive and significant coefficients; male tertiary educational attainments have negative and statistically significant coefficients. - Fertility rates, used as a proxy for children’s effect on women’s decision to work, have negative and marginally significant coefficients. ### Development, urbanization, and labor market efficiency - The polynomial of the log of GDP per capita is statistically significant and generates a U-shaped relationship between female labor force participation (LFP) and level of economic development; Costa Rica is located at the bottom of the U-shape. - Measures of labor market efficiency are positively and significantly related to female LFP. - The share of urban residents has a negative and marginally significant coefficient in cross-country regressions; urbanization can have opposing effects on female LFP (access to services jobs vs. commuting constraints). In contrast, microdata evidence for Costa Rica suggests urbanization may have positive and significant effects on female LFP in Costa Rica specifically. ### Decomposition: Costa Rica versus LA5 - Differences in investment explain a large portion of Costa Rica’s lower female LFP rates compared to LA5. - Investment in telecommunications and transportation contributes significantly to explaining female LFP variation, both in levels and in differences compared to LA5; this is the largest driver of Costa Rica’s female LFP gap with LA5. - Total GDP per capita contributes slightly to Costa Rica’s relatively low female LFP, with Costa Rica featuring at the lowest point of the estimated U-shaped relationship. - The contribution of the residuals to the difference between Costa Rica and LA5 is negative and could reflect elements not captured by the model, including potentially social stigma about women working and cultural elements. ### Policy recommendations - Increase investment in physical infrastructure and in information technology and telecommunications to support female LFP. - Improve implementation capacity to remove constraints on investment; improving implementation could raise female LFP and leverage the large pool of educated women in the country. - Explore measures to support working mothers with children and assess whether low fertility reflects a lack of a developed childcare framework (public, private, or family-based) or other explanations such as cultural norms. - Contextual facts: Costa Rica fertility rate is 1.8 children per woman; Panama fertility rate is 2.5 children per woman; in 2013 Costa Rica’s fertility rate reached the lowest in its history. ### Key cross-country regression statistics (Table 2) - Dependent variable: FLFP rate (FLFP rate under 25 years-old in (6)) - Log GDP per capita coefficients (standard errors in parentheses): - (1) -46.918*** (11.909) - (2) -52.519*** (13.460) - (3) -54.902*** (11.701) - (4) -55.646*** (11.391) - (5) -69.790*** (12.611) - (6) -66.596*** (14.006) - (7) -92.947*** (19.612) - (8) -113.19*** (32.321) - (Log GDP per capita)^2 coefficients (standard errors in parentheses): - (1) 2.511*** (0.641) - (2) 2.757*** (0.719) - (3) 2.845*** (0.619) - (4) 2.944*** (0.605) - (5) 3.638*** (0.672) - (6) 3.525*** (0.758) - (7) 4.623*** (0.987) - (8) 5.963*** (1.857) - Fertility rate per 100: - (1) -1.881 (1.282) - (2) -1.949 (1.352) - (3) -2.267* (1.368) - (4) -2.171 (1.506) - (5) -0.831 (1.728) - (6) -0.695 (1.840) - (7) -2.051 (1.955) - Internet uses: - (2) 0.077 (0.052) - (3) 0.096* (0.049) - (4) 0.110** (0.053) - (5) 0.317*** (0.077) - (7) 0.049 (0.066) - (8) -0.071 (0.107) - Share of urban residents: - (3) -0.102* (0.060) - (4) -0.116 (0.079) - (5) -0.105 (0.090) - (6) -0.135 (0.094) - (8) -0.203 (0.132) - Share of female secondary education: - (4) 0.023 (0.050) - (5) 0.003 (0.064) - (6) 0.086 (0.078) - (8) 0.039 (0.093) - Share of female tertiary education: - (4) 0.182*** (0.061) - (5) 0.066 (0.101) - (6) 0.147** (0.066) - (8) 0.342** (0.146) - Share of male tertiary education: - (4) -0.228*** (0.071) - (5) -0.127 (0.135) - (6) -0.091 (0.060) - (8) -0.416** (0.204) - Labor market efficiency: - (7) 9.012*** (2.053) - (8) 9.424*** (2.499) - Invest to transportation and telecoms: - (7) 1.461* (0.776) - Observations by model: - (1) 4073 - (2) 4069 - (3) 3514 - (4) 3514 - (5) 1789 - (6) 1789 - (7) 592 - (8) 303 - R square by model: - (1) 0.495 - (2) 0.502 - (3) 0.507 - (4) 0.513 - (5) 0.556 - (6) 0.458 - (7) 0.692 - (8) 0.710 *Source: IMF staff estimates and tables reported in the source content.* --- _Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16132.pdf_